All CentOS 6 CTs terminals/ssh is broken after update

[BigWeb]

Hi ,

I havent upgraded proxmox for ages, and did this last week with update to 5.3-12 and latest kernels .

After restart CTs with UBUNTU are functioning fine , however CTs with Centos 6 have these issues :

1. SSH is unavailiable : PuTTy says :

login as: root
root@MyCT's password:
Server refused to allocate pty

SmartTTY says :

SSH PTY request failed (LIBSSH2 returned LIBSSH2_ERROR_CHANNEL_REQUEST_DENIED). Try increasing minimal delay between PTY requests and the extra PTY request delay under Tools->Options->VisualGDB->General->SSH.

2. When i login via main server (pct enter 100), open midnight commander and pres Ctrl-O (to see shell screen) i get error message:

Not an xterm or Linux console;
the panels cannot be toggled.

3. In midnight commander , when entering with "pct enter 100" F1-F8 buttons doe not work , generating
F1 - "11~"
F2 - "12~"
etc...
No different settings in PuTTY help to resolve this .

Please help, as me and my clients can't login drectly and do their work

PS - CTs are on udated CentOS6 - yum update shows no packages to update .

PPS- All these function where functioning just fine before the update

 

[oguz]

I can't seem to reproduce this with our CentOS6 template. Did you reboot your PVE after the kernel upgrades?

Can you post the config from one of these CTs? `pct config CTID` should do it.

 

[BigWeb]

OK, I have ran another upgade to upgrade all packages to final . I also did reboot of main server (PVE) after upgrade to kernel 4.15.18-12-pve, so i have no upgrades availiable in both PVE and CT . and both where rebooted .

Same situation

pct config 320 gives:

arch: amd64
cores: 2
hostname: MYHOSTNAMEMASKED.com
memory: 2000
net0: name=eth0,bridge=vmbr0,gw=X.X.X.X,hwaddr=3E:AE:DA:E9:EF:6B,ip=Y.Y.Y.Y/32,type=veth
onboot: 1
ostype: centos
rootfs: local-zfs:subvol-320-disk-1,size=99999M
swap: 4000

 

[oguz]

Which version of PVE did you use before the upgrade? If you are coming from PVE 3.x, then your containers were most likely OpenVZ containers and not LXC.

Please check the `/etc/fstab` file in the containers. Do they have '/dev/pts' and '/dev/shm' entries?

If that is the case, you will need to adapt the /etc/fstab of the container. See here for the exact instructions: https://pve.proxmox.com/wiki/Convert_OpenVZ_to_LXC#PTY_Allocation in section 'PTY Allocation'.

If you have further problems converting from OpenVZ to LXC, I suggest you to read this page thoroughly for more information.

 

[BigWeb]

Yes theres 2 lines in fstab.
I'll try removing them, however - VPS starts ok, and works ok (in terms of apache/Mysql) (in manual it says remove them if container fails to start).
Anyway - I'll report here later.

 

[BigWeb]

OK, This solved first part of problem the problem indeed !
(Looks like all CentOS 6 containers have this lines in fstab (probably where cloned form one container) )

However the question with F1-F8 buttons stayed :

if in PVE's shell - all works as required .
once i enter container : pct enter XXX - inside it F1-F8 generated (11~, 12~ etc) , while F9 -F12 works fine.
(Tho its not very serious problem, however would be good to know if there's a way of solving this , as again - it was ok before upgrade , so i assume some settings probably got wrong ?)

Would be very gratefull if you could pinpoint me on this as well , and thanks again for fast fix solution.

 

[alexskysilk]

never mind then

It just so happened that I ran into this thread as I was upgrading my test cluster so I spun up a centos 6 ct before upgrading, and then fired it up after the upgrade was complete and rebooted.

the container was accessible via ssh normally. There must be additional variables.
1. How are you routing access to the machine? there is a nat of somekind I presume; can you ssh in successfully inside the network?
2. did anything else change (firewall rules on the router/proxmox/iptables in the container?)

 

[oguz]

once i enter container : pct enter XXX - inside it F1-F8 generated (11~, 12~ etc) , while F9 -F12 works fine.
(Tho its not very serious problem, however would be good to know if there's a way of solving this , as again - it was ok before upgrade , so i assume some settings probably got wrong ?)

I don't understand?